[Ipopt] About the termination conditions of IPOPT

[Hashemi, Reza]

Hello everyone,

 

In the context of optimal control, I am using the IPOPT to solve a huge
optimization problem.

The problem is that it happens very often that the optimal point
computed by the IPOPT has a larger cost than the supplied initial point
(minimization problem) even in the case that the initial point is
acceptably feasible.

I assume the tuning of the options which determine the termination of
IPOPT cause this problem. However, it seems tuning these options for an
ill-conditioned problem is not trivial.

My question is: Is there any method (or option) that enforces the IPOPT
to give a solution which has less cost than the initial point (or at
least as good as the initial point)?

 

Thanks in advance!

Reza

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/ipopt/attachments/20150130/d44ad8e0/attachment.html>